Incidentalomas, unsuspected adrenal tumors that are encountered during radiographic testing done for unrelated causes, are among the most common solid organ tumors in humans. This classification carries a very broad differential diagnosis (Table 12-1) and spans the spectrum of aggressiveness in tumors. Although most incidentalomas are benign and require no intervention, adrenocortical cancer (ACC) is among the most aggressive cancers in humans. Moreover, a significant subset of incidentalomas is hormonally functional, and although these tumors are histologically benign, failure to diagnose and treat patients with these tumors results in increased morbidity and may risk mortality.

With the increasing use of cross-sectional, sonographic, and other imaging technologies, incidentalomas are likely to be diagnosed even more frequently in the future. After they have been identified, these tumors must be characterized as benign or malignant (and if malignant, as primary vs. metastatic) and as hypersecretory or nonhypersecretory. Because the majority of incidentalomas are ultimately managed nonsurgically, their increasing incidence and their follow-up are likely to become an even more important diagnostic, logistic, financial, and therapeutic challenge for individual practitioners and for our health care system. This chapter discusses the prevalence, differential diagnosis, and diagnostic and therapeutic strategies of these fascinating tumors.

First described in the early 1980s,1 the term incidentaloma refers to all clinically unapparent adrenal tumors that are incidentally discovered during imaging studies that are done for other causes. Tumors that are discovered during the initial assessment or follow-up of patients with nonadrenal malignancies are excluded from this definition. Similarly, large adrenal tumors that are symptomatic because of mass effect are also excluded. The reported prevalence of these tumors ranges from 0.35% to 5%2 and is affected by the sensitivity of the radiologic examination that is being used and the inclusion criteria of patients. Thin-cut helical computed tomography (CT) scanning and high-resolution magnetic resonance imaging (MRI) can now identify subcentimeter adrenal lesions that were previously undetectable. Moreover, studies that include many elderly patients or patients with symptoms that can retrospectively be attributed to their adrenal masses report a higher frequency of incidentalomas and have a higher frequency of malignant or metastatic tumors in the adrenal gland.

Similar to the variation seen in the prevalence reported by radiologic tests, autopsy data are also inconsistent. Depending on inclusion criteria, the reported prevalence of adrenal tumors is between 1.05%3 and 32%.4 Autopsies of hypertensive patients reveal an even higher prevalence of adrenal neoplasms, as high as 68%.5 A recent report of pooled autopsy data suggested that the overall prevalence of adrenal adenomas is 5.9% in patients without premortem evidence of adrenal disease or malignancy. This prevalence was age related: although only 0.2% cadavers from patients younger than age 30 years had adrenal adenomas, 6.9% of cadavers from patients older than age 70 years had adrenal adenomas.6

All adrenal tumors, regardless of their radiographic appearance or suspicion for malignancy, should be screened to exclude biochemical hyperfunction. A suggested algorithm is depicted in Figure 12-1.

Pheochromocytoma

Although the incidence of pheochromocytoma in the general population is only 2 to 8 per million,7 these tumors account for 11% to 23% of incidentalomas.2,5,8,9 The classic triad of symptoms attributable to pheochromocytoma is hypertension, headaches, and flushing, but this symptom complex is only present in approximately 40% of patients with an established diagnosis.10,11 Partly because pheochromocytomas may be clinically silent,10,11 up to 75% of pheochromocytomas are not diagnosed until after the patient’s death.12 Undiagnosed, the stimulation of a pheochromocytoma by the induction of anesthesia,13 percutaneous biopsy, or surgery is associated with death rates up to 80%.5,12 For 10% of all pheochromocytomas, the initial manifestation is pheochromocytoma crisis caused by spontaneous hemorrhage or rupture.10

Because the consequences of failure to recognize a pheochromocytoma are so dire, this entity must be definitively established or excluded in any patient who presents with an adrenal mass, irrespective of symptoms. Thus, considerable research has been done to determine which of the multiple available biochemical and radiographic tests most conclusively secures the diagnosis.

A recent multicenter cohort study compared all available biochemical (urine and plasma) tests for the diagnosis of pheochromocytoma.14 The authors concluded that plasma-free metanephrines, with sensitivity of 99% and specificity of 89%, were the best tests for the confirmation or exclusion of pheochromocytoma. Moreover, because the addition of a second diagnostic test did little to improve the diagnostic accuracy in patients with negative plasma metanephrine test results, the authors recommended against the use of confirmatory tests for most cases. This study formed the basis for the recent recommendations made at the National Institutes of Health (NIH) consensus conference.15

Researchers from the Mayo Clinic7 disagree. They point out the role of pretest probability (prevalence) in the determination of the predictive value of a diagnostic test and suggest that for the general population, the routine use of plasma metanephrines would result in an unacceptably high (30%) number of patients with false-positive test results. Therefore, they recommend that this test be reserved only for situations with high pretest probability of pheochromocytoma (patients with a genetic syndrome, family history, or radiologic features) of pheochromocytoma. For the more typical scenario of a patient with a clinical suspicion (based on difficult-to-control hypertension, palpitations, or low-attenuation incidentalomas), they suggested that 24-hour urinary metanephrines and catecholamines, with significantly higher specificity and minimally lower sensitivity than plasma-free metanephrines, be used as the test of choice.

To minimize the number of false-positive results, a threshold of double the upper limit of normal is recommended.16 Moreover, testing should be done under circumstances that avoid artificial elevations in catecholamine levels: the patients should be supine and relaxed, and the blood collection should be through an indwelling catheter rather than a direct venopuncture. Caffeine, nicotine, and acetaminophen are all known to increase catecholamine levels and should be withdrawn for 2 weeks before testing.14

Based on these studies, our practice is to use plasma metanephrines as a screening test only. Because of its better specificity, we confirm the diagnosis using 24-hour urine metanephrine measurements. We recently studied a group of 10 patients with adrenal incidentalomas and borderline elevated urine or plasma metanephrine levels.17 Among this group of completely asymptomatic patients, three (30%) had a pheochromocytoma, highlighting the need for preoperative α-blockade in all patients who have any elevation in their metanephrine levels.

Cushing’s Syndrome

Although the incidence of Cushing’s syndrome (CS) is exceedingly low in the general population,18 studies of specific high-risk subpopulations have revealed an unexpectedly high prevalence of unsuspected CS. In newly diagnosed diabetics; poorly controlled diabetics; and obese patients with hypertension, diabetes, or polycystic ovary syndrome, the frequency of unsuspected CS was 2.0% to 3.3%, 1.0%, and 5.8%, respectively.19–21 Even among patients with hypertension, screening studies have found a 0.5% to 1.0% prevalence of unsuspected CS.22,23

Although a variety of biochemical tests are available for the evaluation of patients with suspected CS, they are all designed to detect aberrations in the normal hypothalamic–pituitary–adrenal axis. In normal individuals with typical sleep–wake cycles, serum adrenocorticotropic hormone (ACTH) and cortisol levels begin to increase in the early morning hours, peak between 7 and 9 AM, and decrease to a nadir for the remainder of the day as long as the patient remains unstressed or asleep. In these patients, the delivery of supraphysiologic doses of glucocorticoids results in suppression of ACTH and cortisol release. On the other hand, CS is characterized by the loss of diurnal variation in ACTH and cortisol release, and patients’ serum cortisol levels remain persistently elevated throughout the day.24,25 Moreover, their cortisol release is autonomous (either because of a primary adrenal tumor or because of an ACTH-secreting mass) and is not suppressed by low-dose glucocorticoid administration.26

Serum Cortisol Levels

Because results of serum cortisol testing are directly affected by patients’ albumin and cortisol-binding globulin (CBG) levels, this test is prone to false-positive and false-negative results under a variety of conditions. Estrogen-containing birth control pills increase CBG levels, and women taking these medications have a 50% chance of a false-positive result.27 They should therefore undergo testing after a 6-week withdrawal period whenever possible.28 On the other hand, malnourished or critically ill patients and patients with the nephrotic syndrome will have falsely decreased serum cortisol levels owing to their low albumin levels.29,30 Furthermore, because of the variation in the degree of hypercortisolism that patients with CS manifest, it is recommended that all testing be duplicated to decrease the likelihood of false-negative results.18

Dexamethasone Suppression Test

A number of published protocols have discussed the performance and interpretation of the dexamethasone suppression test (DST).31–34 The overnight DST is the simplest version of this test and requires the administration of 1 mg of dexamethasone between 11 PM and midnight followed by measurement of serum cortisol between 8 and 9 AM on the following morning. Because patients with CS have variable degrees of responsiveness to dexamethasone, researchers have recommended that a low threshold (postsuppression cortisol level <1.8 μg/dL) be used to enhance the sensitivity of the test.18,34

The 48-hour, 2-mg/day low-dose DST is preferred by some endocrinologists because of its improved specificity over the 1-mg test.18 During this examination, which can also be administered in the outpatient setting, 0.5 mg of dexamethasone is given every 6 hours for 2 days, and a final cortisol level is measured 6 hours after the last dose. As in the low-dose test, the cortisol level should be suppressed to less than 1.8 μg/dL in patients without CS.

Urinary-Free Cortisol Levels

Urinary-free cortisol (UFC) measurements, unlike serum cortisol measurements, are unaffected by patients’ CBG or albumin levels. Measured over a 24-hour period, UFC provides a reliable evaluation of the patient’s total daily cortisol secretion. Falsely low measurements of UFC may occur in patients whose creatinine clearance is less than 60 mL/min,35 and elevated UFC is seen in patients with excessive (>5 L/day) fluid intake36 and in patients with physiologically increased cortisol levels such as during pregnancy and in patients with depression, alcoholism, morbid obesity, or poorly controlled diabetes.18

To optimize the test’s reliability, patients should be carefully instructed to avoid all glucocorticoid-containing medications during the collection period and to avoid excessive fluid intake. The collection should exclude the first morning’s void but include all subsequent voids for the next 24 hours, including the second day’s first morning void. During the test period, the specimen should be refrigerated. The test should be duplicated to increase its accuracy, particularly in children.18

Late Night Salivary Cortisol Level

This test relies on the fact that free serum cortisol rapidly (within a few minutes) equilibrates with salivary cortisol.37 Although many methods can be used to measure salivary cortisol, the enzyme-linked immunosorbent assay (ELISA) and liquid chromatography-mass spectrometry/mass spectrometry (LC-MS/MS) tests are best validated and most commonly used.18 Saliva is collected either passively (by collecting drool into a capillary tube) or actively (by placing a cotton pledget in the mouth while the patient is chewing). With a sensitivity and specificity between 92% and 100% for the detection of CS, late night salivary cortisol has accuracy that is similar to that of UFC.38 This test’s accuracy may be decreased in critically ill patients, patients in whom the circadian rhythm is blunted, elderly men, smokers, and heavy drinkers.18

Aldosteronoma

Primary hyperaldosteronism is the most common cause of secondary hypertension, and it is usually curable by adrenalectomy. Among patients with hypertension, the prevalence of aldosteronoma was estimated to be approximately 1%,39 although more recent reports suggest that the true prevalence of aldosteronoma is at least 10 times higher than initially thought, even among normokalemic patients.40 Among patients with incidentaloma, however, aldosteronoma is a relatively uncommon diagnosis. In a recent series of more than 1000 incidentalomas,41 the prevalence of aldosteronoma was only 1.4%. This low frequency is likely because of patient selection because many aldosteronomas are discovered as part of a workup for hypertension.

Although the hallmark of primary hyperaldosteronism (Conn’s syndrome) is hypertension and hypokalemia, neither finding is necessary for the diagnosis. There are scattered reports of normotensive primary hyperaldosteronism,42 and even Conn himself described a large number of patients with normokalemia in the setting of aldosteronoma.43,44 Nevertheless, the absence of hypertension effectively rules out the presence of an aldosterone-secreting tumor, and most experts recommend no biochemical workup to exclude aldosteronoma for normotensive patients who are found to have an incidentaloma.5,8